Background: Type 2 diabetes mellitus (T2DM) is a condition characterized by insulin resistance and progressive failure of the insulin-secreting beta-cells. Previously considered a disease of adults, it is now becoming increasingly prevalent in children and adolescents. Patients with childhood onset T2DM are at very high risk for diabetes-related morbidity and mortality, due to a longer life-time duration of diabetes, as well as possibly increased rapidity of beta-cell failure. In part, the impairment in insulin secretion in T2DM is caused by beta-cell exhaustion due to a constant, unsuccessful attempt to compensate for the existing insulin resistance. In addition, beta-cell function is affected by glucotoxicity, generating a downward cycle of hyperglycemia leading to decreased insulin secretion, which further worsens hyperglycemia. Results from two recent studies in adults with newly diagnosed T2DM suggest that intensive insulin treatment for 2 weeks may break this cycle, resulting in significant, long-term improvement of beta-cell function. Both reports documented that approximately 50% of patients maintained euglycemia on diet alone at 12-month follow-up. The concept of beta-cell rest, or suppression of insulin release from beta-cells, was originally developed in the context of type 1 diabetes. It was then expanded into the field of T2DM and argues that decreased demand on beta-cells can lead to improvements in insulin secretion and beta-cell viability. Over the past 30 years, results of many in vitro, in vivo and clinical studies support this hypothesis. Study design: In this randomized, controlled trial we will divide adolescents and young adults with T2DM of up to 2 years duration into 2 treatment groups: Group 1 (control arm) will receive conventional therapy for T2DM (metformin plus diet and behavior modification). Group 2 will undergo beta-cell rest using continuous subcutaneous insulin infusion (CSII) for a period of 2 weeks, in addition to conventional therapy (metformin plus diet and behavior modification). The study will last for one year, and patient participation consists in the following: a) 2-month run-in period with metformin, diet and exercise b) 2-week intensive treatment if randomized to group 2 c) Outpatient follow-up visits every 3 months for one year (metformin, diet and exercise will be continued in both groups throughout) The primary outcome will be a comparison of insulin secretion (assessed at one year) in the beta-cell rest group versus the conventional group at the 12-month follow-up time point. Within the beta cell rest protocol we have conducted a pilot project entitled: "Effects of Artificial Sweeteners on Intestinal Glucose Absorption". Most recently, we received approval from the Institutional Review Board for a new protocol (10-DK-0163) in which we will study the metabolic effects of sucralose in greater detail. This protocol is entitled: "Metabolic Effects of Non-nutritive Sweeteners". Background Ingestion of sodas containing artificial sweeteners is common practice in healthy persons as well as patients with metabolic abnormalities. It is generally assumed that glucose metabolism is not altered since these sodas contain no or extremely few calories from carbohydrate. Health concerns rather stem from long-term use (e.g. bladder cancer in association with cyclamate and saccharin). However, recent data obtained from animal studies indicate that sucralose (Splenda) not only activates receptors in lingual taste buds but also in brush cells in the gut, which subsequently activate gut glucose transporters (GLUT2). The intestinal glucose absorption rate is markedly enhanced and can be reversed by co-administration of the selective GLUT2 inhibitor phloretin. Much less is known about intestinal taste receptors in humans. To our knowledge, there are no reports investigating whether similar mechanisms exist in humans. Hypothesis We speculate that ingestion of diet soda before an oral glucose load (OGTT) will lead to faster glucose absorption. In healthy volunteers, who are capable of prompt insulin secretion to compensate for increased glucose absorption, diet soda ingestion will thus induce an earlier and higher insulin response. Subjects with diabetes (either type 1 or type 2), will be unable to mount an appropriate insulin response to the faster glucose absorption, and thus will have higher blood glucose levels. Patients with early (non insulin-requiring) type 2 diabetes are capable of secreting variable, but clinically significant amounts of insulin in response to an oral glucose load. Thus, they may have slightly elevated blood glucose levels due to the faster glucose absorption triggered by the artificial sweeteners followed by a sluggish, late insulin response. In contrast, subjects with type 1 diabetes have much more significant impairment of insulin secretion and therefore even less capability to compensate for increased intestinal glucose uptake. As a result, patients with type 1 diabetes should show a clear increase in blood glucose levels during an OGTT following diet soda ingestion if our hypothesis is correct. Study Design A group of subjects (15 healthy volunteers, 15 subjects with type 2 diabetes and 15 subjects with type 1 diabetes) will undergo two 4-hour oral glucose tolerance tests (OGTT). Immediately before one of the OGTTs, subjects will either drink a sucralose-containing diet soda or an equal volume of sparkling mineral water. The order of the OGTTs (with vs. without diet soda) will be randomized, with subjects serving as their own controls. Results of healthy volunteers participating in this pilot project have been summarized in a manuscript submitted to Diabetes Care. Briefly, glucose excursions were similar following carbonated water and diet soda. Serum insulin levels tended to be higher after diet soda, without statistical significance. GLP-1 peak and area under the curve (AUC) were significantly higher with diet soda (AUC 79.2 50.1 pmol/L180min) vs. carbonated water (AUC 53.5 29.7 pmol/L180min, p = 0.003). We concluded that artificial sweeteners synergize with glucose to enhance GLP-1 release in humans. This increase in GLP-1 secretion may be mediated via stimulation of sweet-taste receptors on L-cells by artificial sweetener. A total of 80 participants have been enrolled in the beta cell rest protocol to date including 11 patients with type 2 diabetes. Six patients with type 2 diabetes have completed the 12 month visit, but it is too early to provide conclusions on the overal effectiveness of the intervention (beta cell rest).